Validation of Diabetes Prediction Scores: Does adding a high risk for depression increase the area under the curve?
Salinero-Fort, M. A.; Mostaza, J.; Lahoz, C.; Cardenas-Valladolid, J.; Iriarte-Campo, V.; Estirado-de Cabo, E.; Garcia-Iglesias, F.; Gonzalez-Alegre, T.; Fernandez-Puntero, B.; Cornejo-del Rio, V.; Sanchez-Arroyo, V.; Sabin-Rodriguez, C.; Lopez-Lopez, S.; Gomez-Campelo, P.; Taulero-Escalera, B.; Rodriguez-Artalejo, F.; San Andres-Rebollo, F. J.; de Burgos-Lunar, C.
Show abstract
BackgroundDiabetes risk scores include age, waist circumference, body mass index, hypertension, use of blood pressure medication, and metabolic and lifestyle variables. Although patients with major depressive disorder have a higher risk of diabetes, none of the diabetes risk scores includes high risk of depression as an additional item. AimTo validate three diabetes risk scores (FINDRISC, DESIR, ADA) in the Spanish population aged >45 years with the aim of predicting diabetes and to test the value of adding high risk of depression, defined as a PHQ-9 questionnaire score [≥]10, to the risk score with the best discriminative performance. MethodsProspective population-based cohort study in Madrid (Spain). FINDRISC, DESIR, ADA, PHQ-9, and OGTT values were measured at baseline. Participants with OGTT <200 mg/dl (n= 1,242) were followed up for a median of 7.3 years using their general practitioners electronic health record (EHR) and telephone contact. Incident diabetes was identified as treatment for diabetes, fasting plasma glucose [≥]126 mg/dl, a new diagnosis in the EHR, or self-reported diagnosis. At the end of the study, the performance of diabetes risk scores, including a modified original FINDRISC score with a new variable for high risk of depression (FINDRISC-MOOD), was assessed. ResultsDuring follow-up, 104 (8.4%; 95% CI, 6.8-9.9) participants developed diabetes, and 185 had a PHQ-9 score [≥]10. The AUROC values were 0.70 (95% CI, 0.67-0.72) for FINDRISC-MOOD and 0.68 (95% CI, 0.65-0.71) for the original FINDRISC. The AUROC for DESIR and ADA were 0.66 (95% CI, 0.63-0.68) and 0.66 (95% CI, 0.63-0.69), respectively. There were no significant differences in the AUROC between FINDRISC-MOOD and the remaining scores. ConclusionFINDRISC-MOOD performed slightly better than the other risk scores, although the differences were not significant. FINDRISC-MOOD could be used to identify the risk of future diabetes.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Association of METS-IR Index with Type 2 Diabetes: A Cross-sectional Analysis of National Health and Nutrition Examination Survey Data from 2009 to 2018 96%
- A cross-sectional questionnaire study: impaired awareness of hypoglycaemia remains prevalent in adults with type 1 diabetes and is associated with the risk of severe hypoglycaemia 95%
- Determinants of glycemic control among persons living with type 2 diabetes mellitus attending a district hospital in Ghana 95%
Similar papers in this journal
- Time In Range, as measured by continuous glucose monitor, as a predictor of microvascular complications in type 2 diabetes– A systemic review 95%
- Comprehensive validation of fasting- and oral glucose tolerance test-based indices of insulin secretion against gold-standard measures 95%
- CREBRF missense variant rs373863828 has both direct and indirect effects on type 2 diabetes and fasting glucose in Polynesians living in Samoa and Aotearoa New Zealand 94%
Similar papers in this journal
- Waist circumference thresholds predicting incident dysglycemia and type 2 diabetes in Black African men and women 96%
- Covid-19 fatality prediction in people with diabetes and prediabetes using a simple score at hospital admission 94%
- A Microbiome-targeting Fiber-enriched Nutritional Formula is Well Tolerated and Improves Quality of Life and Hemoglobin A1c in Type 2 Diabetes: A Double-Blind, Randomized, Placebo-Controlled Trial 94%
Similar papers in this journal
- Using the illness-death model to estimate age- and sex-standardized incidence rates of diabetes in Mexico from 2003 to 2015 95%
- Awareness, knowledge and trust in the Greek authorities towards COVID-19 pandemic: results from the Epirus Health Study cohort 92%
- The association of lifestyle with cardiovascular and all-cause mortality based on machine learning: A Prospective Study from the NHANES 92%
Similar papers in this journal
- The Relationship Between Clinical Profile, Glycemic Parameters And Hypoglycemia In Type 1 Diabetes Pediatric Patients 93%
- Exploring the Association Between Urinary Incontinence and Depression Based on a Series of Large-Scale National Health Studies in Turkiye 93%
- Using machine learning to evaluate the value of genetic liabilities in classification of hypertension within the UK Biobank 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.